| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Pulmonary, Allergy and Critical Care Medicine, The Cleveland Clinic Foundation, Cleveland, OH.
Correspondence to: Mani S. Kavuru, MD, FCCP, The Cleveland Clinic Foundation, 9500 Euclid Ave, Mail Code A-90, Cleveland, OH 44109; e-mail: kavurum{at}ccf.org
Abstract
Pulmonary alveolar proteinosis (PAP) is a rare idiopathic autoimmune lung disease in adults characterized by the accumulation of lipoproteinaceous material within the alveoli of the lung. The natural history of this disease is poorly defined. Current therapy of bilateral whole-lung lavage (WLL) under general anesthesia is invasive and has its limitations. Data suggest that relative granulocyte macrophage colony stimulating factor (GM-CSF) deficiency may be involved in the pathogenesis of this disease. There have been several case series that have described clinical improvement with exogenous GM-CSF therapy in a subset of patients with PAP. We describe the results of a prospective, open-label clinical trial of daily subcutaneous GM-CSF therapy in a group of adult patients with idiopathic PAP. In this series of 25 patients, the largest reported to date, administration of GM-CSF improved oxygenation as assessed by a 10 mm Hg decrease in alveolar-arterial oxygen gradient, as well as improvement in other clinical and quality of life parameters in 12 of 25 patients (48%) with moderate symptomatic disease who completed the trial. In addition, the serum anti–GM-CSF antibody titer correlated with lung disease activity and was a predictor for responsiveness to therapy. These data indicate that subcutaneous GM-CSF therapy is a promising alternative to WLL for symptomatic patients with PAP.
Key Words: antibodies • granulocyte macrophage colony stimulating factor • immunotherapy • pulmonary alveolar proteinosis
Pulmonary alveolar proteinosis (PAP) is an uncommon idiopathic disease characterized by the deposition of extracellular granular lipoproteinaceous material within the lung alveoli. To date, < 500 cases have been reported in the literature.1 There is no specific therapy for PAP. Sequential whole-lung lavage (WLL) is the standard of care.2 WLL is a cumbersome and invasive procedure that requires general anesthesia and double-lumen endotracheal intubation. WLL does not correct the underlying defect in PAP.
Although there is persisting uncertainty about the pathogenesis of PAP, current thinking indicates that idiopathic adult PAP is an autoimmune disorder characterized by circulating anti-granulocyte macrophage colony stimulating factor (GM-CSF) antibodies and dysfunction in GM-CSF signaling, which result in an abnormality in surfactant clearance from the alveoli. Murine gene-targeting studies345678 have shown that mice that are homozygous for a disrupted GM-CSF gene acquire a lung lesion that is histologically similar to PAP, which can be resolved by exogenous GM-CSF or by local expression of GM-CSF by a surfactant protein-C promoter in a transgenic mouse model. The adult human disease is likely not due to GM-CSF gene deletion or receptor abnormality.8
Based on the murine studies, prior to the description of the anti–GM-CSF antibody in patients with PAP by Thomassen et al8 and Kitamura et al,9 Seymour et al10 first reported an adult patient with idiopathic PAP in whom gas exchange improved significantly with administration of subcutaneous GM-CSF, subsequently worsened with discontinuation of the drug, and improved again with readministration of GM-CSF. This preliminary experience was followed by two small open-label prospective phase II studies1112 of GM-CSF in PAP. In addition, there have been several additional case reports1314 describing the efficacy of this therapy. In this article, we summarize the results of a prospective, open-label study of recombinant human GM-CSF therapy in symptomatic moderately severe idiopathic or primary PAP in a consecutive sample of 25 adult patients at a single center. We have described earlier the effect of GM-CSF therapy on the first four PAP patients from our institution.11 We have also described that anti–GM-CSF titer predicts response to GM-CSF therapy in 11 patients and the diagnostic utility of the anti–GM-CSF antibody.1516 This experience represents the largest series to date of a pharmacologic therapy for PAP.
Materials and Methods
The study was a prospective, open-label clinical trial of daily GM-CSF therapy in a group of patients with idiopathic PAP. The study was approved by the Institutional Review Board at the Cleveland Clinic Foundation, and informed consent was obtained from all patients prior to enrollment. A diagnosis of idiopathic PAP was confirmed by transbronchial lung biopsy (n = 8) or open-lung biopsy (n = 17) in all patients. All patients had pretreatment serum assayed for GM-CSF–neutralizing antibody as previously published.15 Approximately 10 subjects were initially planned for inclusion in the study for a treatment duration of 10 weeks starting from March 1998. Preliminary efficacy and safety data indicated that extending the duration of therapy to 12 months was potentially beneficial. Hence, the total enrollment was increased to 20 subjects in October 1999 as well as the duration of therapy to 12 months. Based on additional data on the first 20 patients, the Institutional Review Board approved an increase to 25 patients in April 2001.
The eligibility criteria included the following: (1) age 
18 years; (2) primary idiopathic PAP; and (3) moderate disease, as defined by the presence of symptoms attributable to PAP (eg, dyspnea, cough), the need for supplemental oxygen 
6 L/min with PaO2 > 55 mm Hg at rest, and diffuse pulmonary infiltrates on a chest radiograph (CXR). Patients with a history of two or more lavages in the prior 4 months of presentation were eligible to participate in the trial 3 months after the last WLL for severe PAP exacerbation. Exclusion criteria included the following: (1) PAP resulting from another condition (eg, myeloproliferative disorder or leukemia, occupational exposure to silica, HIV disease, or respiratory infections); (2) increased risk of side effects with GM-CSF therapy (eg, rheumatoid arthritis, immune thrombocytopenia, autoimmune thyroiditis); (3) significant cardiac, renal (eg, creatinine 2 mg/dL), or liver disease (ie, hepatocellular enzyme levels exceeding three times normal); (4) severe PAP (defined as PaO2 < 55 mm Hg while receiving 6 L/min of oxygen, and those requiring inpatient care and more urgent therapy with bilateral WLL); (5) pregnant and lactating women; and (6) active associated respiratory infection by history, physical examination, and CXR.
Dosage and Dose Escalation
Patients were treated with recombinant human yeast-derived GM-CSF (Leukine; formerly Immunex Corporation and now Berlex; Seattle, WA) administered subcutaneously once daily for 3 months. Patients were trained to self-administer daily subcutaneous therapy. The starting dose was 250 µg/d, which was progressively increased to 5 µg/kg/d for the second month of study, and to 9 µg/kg/d for the third month. After the patient received therapy for 3 months, if the clinical response was still suboptimal and the patient was tolerating therapy, the patient underwent further dose escalation from 9 to 18 µg/kg/d (12 µg/kg/d at 3 months, 15 µg/kg/d at 4 months, and 18 µg/kg/d at 5 months). If the patient had an adequate response based on above-mentioned criteria, that dose was continued from 3 to 12 months.
Monitoring During Therapy
The first dose was administered in the clinic, and patients were observed for 2 h. Injection sites were rotated between the anterior abdominal wall and upper thighs. Outpatient follow-up occurred at 2 weeks and months 1, 2, 3, 4, 5, and 6 after initiation of therapy. During follow-up, a CXR, complete pulmonary function studies, a 6-min walk test, a dyspnea questionnaire, and a Short Form-36 (SF-36) quality of life questionnaire were obtained. A CBC with differential was obtained on a weekly basis for the first month and then biweekly on the study medication. After the 6-month visit, the patients were seen at 9 months and 12 months. Patients were examined by a physician for signs of any drug-related toxicity during follow-up visits. Follow-up telephone calls were also made by the research coordinator at weeks 30, 42, and 80.
Parameters for Study Discontinuation
The study was discontinued if patients had one of the following: (1) progression of the underlying PAP leading to deterioration of the respiratory status requiring hospitalization for WLL (assessed based on symptoms, physical examination, CXR, and oxygenation); (2) serious allergic or anaphylactic reaction with the first dose of GM-CSF or during subsequent therapy; (3) persistent significant laboratory abnormalities including, but not limited to, WBC count > 50,000/µL, absolute neutrophil count > 20,000/µL, platelet count > 500,000/µL, serum creatinine > 2 mg/dL, and liver enzymes more than three times normal; or (4) refractory "capillary leak syndrome" not responsive to lowering the dose and/or diuretic therapy. Capillary leak syndrome was clinically defined as excessive fluid retention, weight gain, peripheral edema, and pleural or pericardial effusion.
Study Outcomes
The primary end point was an improvement in oxygenation as assessed by a 10 mm Hg decrease in the room air alveolar-arterial oxygen gradient (P[A-a]O2). For the patients who received GM-CSF therapy after hospital admission for bilateral lavage (history of two or more WLLs in the prior 4 months), the primary outcome variable was the need for subsequent therapeutic lavage or further hospital admission days (rather than oxygenation). Secondary end points were as follows: (1) improvement in symptoms as assessed by a dyspnea questionnaire, pulmonary function tests, 6-min walk distance with oxygen saturation, and CXR abnormalities; (2) overall tolerability of therapy; (3) requirement for WLL and sustainability of a response as assessed by follow-up telephone interview; and (4) quality of life as measured by the SF-36 survey.
Radiographic Grading
Single frontal CXRs were graded as to the presence and extent of lung opacification using a visual scoring system similar to one described by Remy-Jardin et al.17 Scores were determined for degree of opacification, extent of opacification, and severity of opacification.
The CXRs (n = 21) were interpreted blindly by our radiologist (M.M.), without the knowledge of clinical, laboratory, or pulmonary function test results, or whether the radiographs were before or after treatment. The presence and degree of lung opacification were coded by means of a 3-point scale for each lung (mild ground-glass opacity, 1 point; moderate ground-glass opacity, 2 points; consolidation, 3 points); this was referred to as the opacity score. The extent of lung opacification was estimated by using a 4-point scale (opacities involving < 25% of each lung were scored as 1 point; 25 to 50%, 2 points; 50 to 75%, 3 points; and > 75%, 4 points), referred to as the extent score. We added the scores for each lung to obtain total scores for opacity and extent; possible opacity scores range from 0 to 6, and possible extent scores range from 0 to 8. An overall severity score was calculated for lung opacification by multiplying the opacity and extent scores for each lung and then taking the sum of the two values; possible severity scores range from 0 to 24.
Statistical Analysis
The two groups were compared on the normally distributed variables using Student t test and on the nonnormally distributed variables using Wilcoxon rank-sum test. The responders and nonresponders were compared on the difference between follow-up and baseline variables of interest (PaO2, P(A-a)O2, diffusing capacity, total lung diffusion capacity/alveolar volume, total lung capacity, FVC, 6-min walk distance, WBC count, lactate dehydrogenase [LDH], and the duration and maximum dose of therapy).
Results
Twenty-seven patients with adult idiopathic PAP were screened, and 25 patients were enrolled in the trial. Figure 1 is a flow chart showing the disposition of all screened patients. Circulating anti–GM-CSF antibodies were present in all patients, measured in pretreatment sera as previously described. Among the 25 patients enrolled in the trial, 18 were men and 7 were women. The demographics and baseline characteristics are shown in Table 1 . Of the 25 patients enrolled, 4 patients did not complete the study and were not evaluable. One patient withdrew due to arthralgia, one patient became pregnant and was withdrawn, one patient refused follow-up care, and one patient died of acute hypoxemic respiratory and multiorgan failure.
|
|
10 mm Hg, satisfying the a priori criteria of response to therapy. Analysis of the data on an intent-to-treat basis revealed that 12 of 25 patients (48%) were clinical responders to GM-CSF therapy. Changes of PaO2, P(A-a)O2, diffusing capacity, total lung capacity, and 6-min walk distance of the responders were significantly higher than nonresponders (Fig 2
). The WBC count (Fig 3
), LDH, FVC, and the maximum dose and duration of therapy between the responders and nonresponders were not statistically significant. Table 3
summarizes descriptive statistics for the SF-36 survey at baseline, 6 months, and the difference for the eight scales on a range of values of 0 to 100. Patients improved significantly from baseline to 6 months for all SF-36 measures except bodily pain. We obtained baseline and follow-up anti–GM-CSF titer measurements in 17 patients (8 responders and 9 nonresponders). Figure 4
reveals the baseline and follow-up anti–GM-CSF titers among the responders and nonresponders. The anti–GM-CSF titer decreased in the responders; in the nonresponders, the titer remained the same. Figure 5
shows the posteroanterior view of a representative CXR of a responder at baseline and after 6 months of GM-CSF therapy. CXR scores at baseline and at the end of therapy are shown in Table 4
.
|
|
|
|
|
|
|
Toxicity
Overall, the drug was well tolerated. Table 5
lists the adverse events noted in the study participants. The majority of the side effects were minor and included injection-site edema, erythema, and malaise. Shortness of breath was noted in 10 patients. It is unclear whether the shortness of breath was related to PAP or the GM-CSF therapy. There were no bone marrow or lung toxicities.
|
The primary finding from this open-label study was that administration of GM-CSF improved oxygenation as assessed by a decrease in P(A-a)O2 as well as other clinical variables in 12 of 25 adult patients (48%) with idiopathic PAP with moderately symptomatic disease. Response to GM-CSF conferred a meaningful benefit clinically to the patients with PAP, with improvements in symptomatology (both dyspnea as well as overall quality of life), reduction in supplemental oxygen requirement, frequency of WLL, and also an increase in 6-min walk distance. The laboratory and clinical improvements also correlate with radiographic improvements. Some of the benefits, such as decrease in supplemental oxygen and WLL frequency, were sustained on long-term follow-up (mean duration, 39 months).
The first description of the efficacy of GM-CSF in acquired PAP was by Seymour et al10 in 1996 as a case report. In a continuation of this, in a multinational study by Seymour et al,12 14 patients were treated with 5 µg/kg/d of subcutaneous GM-CSF for 6 to 12 weeks. Patients who did not respond to 5 µg/kg/d of GM-CSF underwent stepwise dose escalation to 20 µg/kg/d, and responding patients were retreated at disease recurrence. Five of 14 patients (36%) showed a response to initial therapy, and 1 of 4 patients responded to a dose escalation of 20 µg/kg/d, and the response lasted a median duration of 39 weeks. Retreatment reproduced the response. GM-CSF–induced eosinophilia was the only treatment-related factor predictive of response. Twelve of 12 patients who had pretreatment sera available were positive for anti–GM-CSF antibodies.
The improvement in clinical parameters was not apparent until 8 to 12 weeks after initiation of therapy in our study. This was similar to the study by Seymour et al,12 in which improvement in P(A-a)O2 was not evident until 4 to 6 weeks, with maximal improvement achieved at approximately 6 to 10 weeks of therapy. These observed delays are consistent with published data181920 that GM-CSF is required for immature precursor cells to be recruited to the lung and stimulated to differentiate into functional alveolar macrophages and, hence, help in surfactant clearance. In a trial21 of aerosolized GM-CSF in GM-CSF–deficient mice, improvement of the lung histology was first evident only after 4 to 5 weeks of therapy.
One observation from our study is that several patients who were receiving GM-CSF treatment noted a response only after prolonged therapy (from 6 to 12 months). Four patients received therapy for 48 weeks, and two patients received therapy for 52 weeks. This is in contrast to the study by Seymour et al,12 in which only seven patients received treatment beyond 10 to 12 weeks and one patient for 26 weeks. Our study suggests that prolonged duration of therapy may be necessary in a subgroup of PAP patients to demonstrate clinical improvement. Whether there are any features that would identify a particular group of patients that would benefit from prolonged therapy is still unclear.
The total (peripheral) WBC count was unchanged with GM-CSF therapy both in responders and nonresponders to therapy. This observation was also noted by Seymour et al.12 We did not find any increase in the total circulating WBC count or eosinophilia with GM-CSF therapy. This is at variance with the study by Seymour et al12 and the case report by Barraclough and Gillies,13 which showed treatment-related eosinophilia was a predictor of response. All patients in our study had anti–GM-CSF antibodies, and it is possible that the antibodies blunt the hematopoietic response to exogenously administered GM-CSF. Two of the patients in the responder group in our study had increases in WBC count with GM-CSF therapy. This is similar to the observation by Seymour et al,22 in which three of their patients with PAP had modest neutrophilia with escalating doses of GM-CSF therapy, suggesting that the responsiveness of the WBC count to GM-CSF therapy in PAP patients was impaired but not absent.22 Two plausible mechanisms for this observation in our patients are that both of these patients received escalating doses of GM-CSF for 26 weeks and 48 weeks, respectively, which may have overcome the inhibitory effect of the anti–GM-CSF antibody. The second mechanism could be that these two patients had a relatively lower anti–GM-CSF titer (1:3200 and 1:6400, respectively) compared to the other responders, which may have made them more susceptible to the stimulating effects of GM-CSF.
We have previously reported that the GM-CSF antibody titer is a predictor of response to GM-CSF therapy.1516 In 11 patients who completed the GM-CSF open-label clinical trial, it was demonstrated that the level of anti–GM-CSF titer is higher in patients with active disease and is lower in patients in remission. The anti–GM-CSF titer at baseline was lower and decreased further in responders to GM-CSF therapy. GM-CSF therapy did not induce or further increase the anti–GM-CSF titer. The data from this study indicate that the anti–GM-CSF titer correlates with PAP disease activity and also the response to therapy. There is also a case report14 of clinically successful treatment of PAP with GM-CSF that was associated with a profound reduction in GM-CSF–neutralizing autoantibodies (both serum and BAL) and improvement in alveolar macrophage morphology and function.
On the contrary, Seymour et al23 analyzed serum of anti–GM-CSF antibodies, surfactant protein-A, surfactant protein-B, LDH, P(A-a)O2, vital capacity, and carbon monoxide transfer factor before recombinant human GM-CSF treatment and every 2 weeks thereafter in 14 patients with alveolar proteinosis. The pretreatment anti–GM-CSF antibody titer did not differ according to therapeutic response. The anti–GM-CSF antibody concentration fell in all patients during their initial treatment with 5 µg/kg/d of recombinant human GM-CSF, but the change in antibody concentration did not differ between responders and nonresponders. A similar pattern was seen in the four evaluable patients who underwent either dose escalation or retreatment after a prior response to recombinant human GM-CSF. Following cessation of recombinant human GM-CSF therapy, sequential assessment in three patients for whom data were available revealed no clear pattern of anti–GM-CSF antibody concentration. The reasons why the anti–GM-CSF antibody concentration did not reflect disease severity or predict response to GM-CSF treatment in this study remain unclear.
The variable natural history of PAP with its possible "spontaneous resolution" makes it difficult to evaluate the effectiveness of GM-CSF therapy. Seymour and Presneill,1 in a review of the published literature, identified only 24 instances of spontaneous resolution objectively documented by CXR or blood gas levels among 303 published cases (7.9%). This suggests that spontaneous resolution of PAP is infrequent, thereby minimizing the possibility that spontaneous resolution was an explanation for the improvements noted in our PAP patients. The other factor that would argue in favor of a treatment effect for the observed clinical response is that many of these patients required frequent WLL for symptom relief before enrolling in our study, suggesting that they did not have many quiescent periods in their disease activity.
GM-CSF therapy was well tolerated, with minor side effects such as local skin reactions. The mean duration of follow-up was 39 months, and we did not see toxicity such as pulmonary fibrosis or bone marrow suppression with this therapy. The duration of follow-up may not be adequate to rule out long-term pulmonary or hematologic toxicity. Long-term follow-up of PAP patients treated with GM-CSF is ongoing.
All patients in our study had pretreatment anti–GM-CSF antibody, indicating an important diagnostic role for this antibody in the noninvasive diagnosis of PAP. In a study16 conducted at our institution, all adult patients (n = 40) with idiopathic PAP were found to have systemic and localized antibodies against GM-CSF, and the autoantibody response was specific for GM-CSF. The lowest serum end-titer for the PAP population was 1:400, and the highest titer was 1:25,600. None of the healthy control subjects had end-titers > 1:10, and two of other pulmonary disease control subjects (beryllium lung disease) had end-titers at 1:100. The sensitivity of the serum anti–GM-CSF assay was 100%, and the specificity improved to 100% (using a cut-off end-titer of 1:400). BAL from 20 of 20 patients with PAP had end-titers of anti–GM-CSF 1:100, while there was no detectable anti–GM-CSF in the BAL of healthy or disease control subjects (patients with beryllium or asthma).16
Questions remain in the optimal management of idiopathic PAP. Prospective data from an international registry are required to establish the natural history of PAP as well as a standardized approach to the technique and timing of conventional therapy with WLL. The exact mechanisms of how GM-CSF deficiency leads to an excess of alveolar surfactant, and the catabolic defect in the alveolar macrophage and/or type II epithelial cell needs to be elucidated to identify other potential therapeutic targets for GM-CSF signaling. GM-CSF therapy should be considered as a viable alternative pending more definitive trials establishing the optimal dose, duration, and route of administration (systemic vs aerosol). Features predicting response and whether there is a daily cumulative dose threshold for response, and long-term safety require further research.
To summarize, GM-CSF therapy improved respiratory disease in nearly one half of our patients. Typically, the response to GM-CSF therapy required 8 to 12 weeks and was not associated with peripheral leukocytosis in PAP. All of our patients had anti–GM-CSF antibodies, which are useful in the noninvasive diagnosis of PAP. Our study shows that the serum anti–GM-CSF antibody titer correlated with lung disease activity and is a predictor for responsiveness to therapy, thereby suggesting a potential option in monitoring during therapy. GM-CSF therapy may be an alternative to WLL (which is a cumbersome, anesthesia-requiring invasive procedure that does not correct the underlying GM-CSF defect in PAP patients) and is well tolerated.
Footnotes
Abbreviations: CXR = chest radiograph; GM-CSF = granulocyte macrophage colony stimulating factor; LDH = lactate dehydrogenase; P(A-a)O2 = alveolar-arterial oxygen gradient; PAP = pulmonary alveolar proteinosis; SF-36 = Short Form-36; WLL = whole-lung lavage
Dr. Venkateshiah is presently at Creighton University, Omaha, NE.
This study was supported by a research grant from Berlex, formerly Immunex Corporation, Seattle, WA.
Received for publication July 25, 2005. Accepted for publication January 17, 2006.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
